Geographical Coverage

• CEEMID started as the Central and Eastern European Music Industry Databases, and one of its aims was to bring these seemingly data-poor areas on an equal footing with their competitors, partners and users on the EU Single Market. We have always focused on finding CEE data and compare it with preferably EU-28 data.

• CEEMID is not focusing on the CEE in data collection, on the contrary: it collected data in data rich Western and Nordic countries, and out of necessity created the data on less covered CEE (and often, Southern) member states.

• CEEMID as a default covers the EU28 on national, and in some cases, NUTS2 regional levels. However, CEEMID makes efforts to include EEA countries Norway, Lichtenstein, Iceland, Switzerland; future and current candidate countries Albania, North Macedonia, Serbia, Turkey; and neighborhood countries Armenia, Ukraine, Georgia.

Data Coverage & Quality

• CEEMID like statistical authorities usually overcollects data. There are two reasons for this: often adding elements, questions to a data collection procedure is almost costless, because the fixed costs are high. Also, if data later will be needed, it is not possible to collect 2013 missing data in 2019.

• The Open Data regime makes data abundant, and often free. However, this data is often semi-processed at most and not tidy, which means that very significant resources are needed to make it usable. This data processing know-how is often missing from microenterprise-sized music industry stakeholders.

• We believe that the often cited example of the European Audiovisual Observatory is not very good, because this observatory was born before the EU, automated data processing and distribution, and the open data regime of the EU. CEEMID currently offers much higher data quality than the European Audiovisual Observatory (website) because it was designed under these circumstances. Creating the European Music Observatory, if technological and regulatory improvements are considered, are a much cheaper, simpler mercerize than the audiovisual observatory’s example would suggest.

Data requirements

• CEEMID would like to transfer its know-how on data availability, data quality and data requirements to the Feasibility Study and to the future Observatory, or make it public for any interested party.

• CEEMID adopted the Standard U.S. / EU data mapping of the music industry for emerging markets (see ⯈ Pillar 1 - our mapping in the entire value chain of the music industry), too, and provides a far more comprehensive data map, i.e. the US/EU standard Three Income Stream Model for collection, processing and creation of music industry indicators than the proposed Four Pillar Model.

• CEEMID understands how to acquire ‘big data’ and other important data via APIs using open source software. We can release the open source code, or the acquired data after consultation with API owners as open data (see ⯈ use case in Pillar 2 - Music diversity and circulation).

• CEEMID can provide know-how on the best practices of following ESSNet Culture guidelines - large pdf on the creation of music indicators (see ⯈ use case in Pillar 3: Music & Society),

• CEEMID can provide the know-how how to adhere to the IFRS Fair Value Measurements principles that are essential to royalty pricing, benchmarking or factual grant evaluation premises

• CEEMID follows the jurisprudence of the Court of Justice of the EU to provide stakeholders with indicators that are admissible as economic evidence, for example in pricing and royalty negotiation disputes, or damage claims (see ⯈ Use case in Pillar 4: Innovation).

Open Data, Re-Use of Public Sector Information

The re-use of public sector information and various open data initiatives made it possible that we integrated many existing, publicly funded, EU-wide data collections without legal hurdles and with very little cost into thousands of indicators that cover all Europe in Pillars 1-3.

• inflation data is made from a vast array of price data collected in EU member states; CEEMID re-uses the price data of cultural goods and services.

• national accounts data is mainly created from the tax and mandatory statistical filings of creative industries; CEEMID uses

• cultural access and participation data are sometimes included in EU research programs,

• music education and training data.

Re-Use of PSI Example: Who Pays for Music on Mobile?

The following indicator, which can be expressed for various target groups, is created solely from available Eurostat data.

Creating Better Statistics

CEEMID has been doing what Eurostat is doing, often years earlier: revisiting newly available open data sources, and creating new regional and national data products.

⎘ Zoom into the presentation online ⎘ read more: Creating better national cultural statistics with Eurobarometer datasets and ESSNet-Culture technical recommendations

Data Integration Instead of Data Collection

• Due to the effective EU Open Data regime, the bulk of CEEMID is mapped & integrated data from existing public sector data collections. The really valuable parts are, however, proprietary and business data elements in combination with and in integration with public sector data.

• We do not redistribute confidential industry data, such as IFPI or CISAC data, however, we fully very comprehensively mapped that data. The metadata and CEEMID software code allows users, for example, the national member of these organizations, to immediately create clean, euro or national currency based indicators jointly with thousands of reliable open statistical and other data.

• We do harvest hundreds of statistical indicators, but we do not only compile public statistics. We access raw statistical data, and create new statistics that is not available from Eurostat.

• We not only clean and tidy public statistical tables, but we integrate them with approximated, forecasted and backcasted values for a more timely and complete international comparison.

Reproducible Research

• CEEMID operates on reproducible research standards. This means that CEEMID’s tables, charts, models can be refreshed from the original source automatically. New annual, quarterly data is always immediately added. Data corrections of public sector bodies is always up to date.

• CEEMID automatically makes currency translations (from dollar to euro, or euro to national currency units), unit conversions (euros, thousands of euros, and millions of euros), and other common transformations (per capita values, full-time equivalent values). Because small music industry stakeholders do not employ data scientists, this saves literally many hundreds of hours of work.

• CEEMID’s reproducible research principle is a roboust solution to lack of data ownership in case of proprietary data. Whenever the original data is not publicly, the inclusion of such data is immediate whenever the data owner approves it.

• CEEMID’s reproducible research principle is a guarantee of higher data quality than with the product of the European Audiovisual Observatory.

Open Source

• CEEMID uses the statistical programming language R, which is often used by national statistical services.

• CEEMID is using only open source software, which keeps or innovative, Pillar 4 products reproducible, re-usable without further costs.

• CEEMID itself contributes to rOpenGov, and releases some of its own software code open sources. We would like to find a public grant to release all ⯈ Pillar 1-3 data collection, cleaning and integration code and an open source license, so that music stakeholders, or the future European Music Observatory can immediately benefit from it. [This is less user friendly, and may be costly because of vast documentation needs.]

• Another solution would be the creation of an Open Data web application, which would not release the software code, but all the data for free. [This is more user-friendly]

• In either case, already existing data pioneers could use thousands of European indicators for the ⯈ Pillar 4 innovative products and serves.

Quality Issues

A large part of the European Music Observatory data coverage is already available, but not well processed. One of the important causes of very low data utilization is that often very significant resources are needed to process the data.

• Currency translations: we always use the correct exchange rates and translation methods among global (dollar), EU (euro) and national (various currency data). We quote growth on a euro/euro, dollar/dollar, national currency unit / national currency unit basis.

• Unit conversions: making sure that euros, thousands of euros, millions of euros, revenue per capita or revenue per thousand people is always consistently used.

• Regional data: In the past 20 years, intra-EU regions changed 5 times. We re-compute regional statistical indicators to follow the changes of national regions to allow a better time-series, panel or cross-sectional analysis of a wide range of indicators on subnational level.

• Tidy data tables: we use the tidy data concept from statistical software design and data science to make sure that our thousands of indicators can be combined. Joining ill-formatted data usually takes up to 80% of user work hours to correct.

Macro-economic patterns and trends

CEEMID uses and re-uses national accounts data (mainly derived from tax returns in member states), financial accounts data, industry collections, own collection, various semi-finished statistical products, and micro-data accessed via the OpenData regime.

• The data availability is very good in Europe. However, it took 5 years for CEEMID to map sources and write programs to automate the collections of thousands of national level indicators.

• This year we started to the creation of subnational and city-level indicators, too.

Trends

Understanding trends requires strong econometric skills and long data series. Many CEEMID indicators have 20 years or longer history, and they are forecasted programatically up to 5 years ahead.

• We believe that one of the biggest economic challenges of the live music industry is seasonality, which creates huge financial risk. It makes capacity utilization of the cultural infrastructure and venues extremely challenging, especially farther away from the big European metropolitan centres.

• We are also experimenting with short-term forecasting on ‘big data’ sources, for example, by comparing actual concert ticket sales with Google search data for concerts.

Value-chain mapping

we mapped the EU music industry based on the Three Income Streams model (pdf), a standard US model adopted for Europe by the JRC Institute for Prospective Technological Studies.

This example is taken from the Mapping the Croatian Music Industry.

Our mapping is necessary to design data integration and primary data collection schemes. CEEMID contains ⯆ transaction quantity and price data from all mapped parts of the value chain.

Collecting Data From the Value Chain

With the help of about 100 stakeholders, we created the largest database of music professional skills, royalty earning assets, working practices and earnings.

Because the average enterprise size is less than 2 people and most musicians, technicians and managers work as freelancers, it is extremely important to collect data from all parts of the value chain. The microenterprises are usually exempted from mandatory statistical data filings, and make simplified tax returns and financial reports. Without surveying them individually, their economic activities remain hidden.

Legal models

In the Legal models we believe that the comparative legal analysis requires so different data, or rather information acquisition and integration strategies that we concentrate only on numeric, measurable information, such as tax levels and tax burdens. While our rich data and comprehensive understanding of the value chain was used to advocate changes in laws and model contracts, this may be the only aspect of the proposed Four Pillars that is not well covered in CEEMID.

⯇ Pillar 1 Review (go back) to Pillar 1 or continue with ⯈ Pillar 2 Music diversity and circulation.

Cross-border circulation of works

CEEMID is working with author/publishing-side , producer/performer/recording-side and live music stakeholders to measure the cross-border circulation of works.

The music industry was among the first industries that went almost entirely digital and where almost all sales were robotized. The industry is characterized by vast amounts of digital data that are exploited by AI algorithms.

Data standardization is just catching up with this extraordinary change, and we are constantly building new tools to automate data collection and integration for all three income streams, i.e. publishing/authors, recording i.e. performers/producers, and the live music stream. We believe that our experience is could be extremely useful in designing and creating a feasible European Music Observatory.

Cross-border mobility of artists and professionals

CEEMID has been surveying musicians about their tour destinations since 2015.

• In several countries our musician surveys are fully representative of the full-time and part-time musician populations, and we have representative touring information.

• We are searching for new ways to reconcile this data with collective management data, because author societies technically should have information about foreign tours.

Cultural diversity aspects

CEEMID has been measuring with surveys the use of various languages in live music, and we are experimenting with various representative ways with collective management societies to measure language use in radio and streaming channels.

• We are currently working on a large case study to understand the effect of national radio quotas, and what can be there lessons for better public broadcaster programming and in streaming.

• We see an extremely fast shrinking market share of domestic and small language works. Because most music is sold by AI algorithms in the world (both in streaming services, and increasingly in radios and background music), small country repertoires are extremely disadvantages, as they do not have large enough datasets and local resources to train better algorithms.

Education, training, personal development

We have been collecting education and training data, and needs data from musicians, technicians and managers since 2015. Our CAP surveys (see ⯆ next slides) contain audience music education variables.

In our grant assessment workshops and surveys, music education and professional development gets the top priority from musicians, technicians and managers. Because about 95% of the industry works in microenterprises, and the rest in small enterprises, strategic HR functions, such as skills assessment, life-long learning are missing in most countries and can be designed only by national stakeholders.

Audiences

The most important revenue source of CEEMID is pricing music and setting royalty levels. According to the international copyright law, the price of music should be based in many uses on the value in use.

• We are collecting data on how music is built into the value chain of restaurants, shops, hotels, which includes data on how people perceive music in this scenes of public performance.

• We are collecting data on how people use music from licensed, exempted (home copying) and illegal sources.

• Data availability is not bad, but in some member states data is available only with little frequency. The data availability is not enough in many countries to find accurate answers to current issues, for example, to understand the actual value transfer from rightsholders to media platforms. In these cases, we are designing and implementing new data collection following the technical guidelines of ESSNet Culture.

• ⯆ Examples of pan-European concert demography data below

Use Case: Audience Demography

In most countries, Europeans visit less and less concerts as they are ageing. In all countries it is starting from the teenager group of 15-18 at 2-3 concerts per year. The visiting frequency drops to 1 concerts in every 2-3 years in the 40-45 age group. Especially in the CEE, this results in a very young concert and poor audience because the higher earning middle-aged people are absent from the venues.

Music and Society

Music and society (not-for-profit sector, associations, social inclusion, amateur music, heritage)

• CEEMID’s music professional surveys have been collecting membership data in associations and non-profit organizations, such as collective management societies since 2014.

• CEEMID’s CAP surveys, and our re-used surveys (see for example ⯇ previous slide on the survey, and a use case in ⯈ next Pillar 4 technological evolution - A.I.) contain indicators on amateur music practices in the general public in all 28 EU member states.

Technological Evolution

CEEMID was created with the aim to provide state-of-art AI/machine learning, analytics products in the geographical areas and industry areas where it is most needed: in less developed EU countries and for the majority of artists who are not represented by labels and publishers with such technologies.

The example shows a comparison of the Czech and general EU audience model which predicts who will likely visit at least one concert in the coming year.

Future Business Models

Future Business Models I: distribution platforms

CEEMID is constantly seeking ways to access DSP and other industry data from application program interfaces (APIs) and follows ESSNet best practices to use big data sources.

The following chart that covers the streaming quantity growth of a typical song in 20 countries was created as an experimental

Future Business Models II: fair remuneration

CEEMID has used all recognized music / intellectual property valuation methods in the music industry.

• We use the ‘music comparators’ model to evaluate to value or private copying, the euro value of the value transfer to media platforms, and to assess the royalties paid by radios and television stations.

• We use ‘hedonic price models’ to valuate various uses, mainly background music uses in restaurants, hotels and shops.

• We use ‘econometric models’ to objectively justify social, economic and cultural differences that must be taken into consideration when pricing music, following the jurisprudence of the Court of the European Union, for example, in the AKKA/LAA vs Konkurences padome case and the Léčebné lázně Mariánské Lázně v OSA case.

• We use hypothetical evaluations where no market data is available, for example, to assess the viability of new products, services, or to make reality checks on the other models.

Complex indicators to establish the value of PCR and value gap 1: Volumes

CEEMID compiles indicators from European CAP surveys accessed via the Open Data regime and creates new CAP surveys following the technical guidelines of ESSNet Culture.

The following example is a randomized example from one of our actual surveys about how people listen to music at home, while they travel and at work. These indicators are essential to use any market comparator models, and they are often required in royalty disputes.

Complex indicators to establish the value of PCR and value gap 2: Prices

CEEMID compiles proprietary data that it never owns, via its data integration model and metadata based on mapping industry sources to compare the actual use of music with actual royalty payments. When such information is not available, we rely on our CEEMID Music Professional Surveys, which contain the earnings data of thousands of European musicians.

The following example is a randomized example that resembles the true values. The unknown quantity is the value of not licensed music, i.e. private copying and value transferred to YouTube or other media platforms. This is a valuation method used by IFPI. The CEEMID market comparator model is a more detailed version of the valuation model created for IFPI by PWC.

Complex indicators to establish the value of PCR and value gap 3: Compensation Scenarios

The following, again randomized example shows suggested private copying remuneration and value gap correction levels with two scenarios with different input parameters. The appropriate level of compensation and target revenue can be compared with actual royalty income.

Music stakeholders need such data when they are in royalty disputes, in litigation, or when they are making regulatory impact assessments. We believe that this information is critical for the correct transposition of the new DSM Directive, too. This example, while eventually results in four values : PCR compensation euro value A, B, and target media platform revenue A, B, relies on the use of dozens of indicators which comply with ESSNet Culture statistical technical recommendations and IFRS Fair Valuation principles. (YouTube is only highlighted because it is far the largest platform, and we emphasis that while we have made real calculations, this is just a randomized number example and not the basis of any claim against YouTube.)

Future Business Models III: authors’ rights collection mechanisms

• CEEMID’s most important users are author’s rights and neighboring rights associations, who need to improve their monitoring, collection and pricing.

New Policies & Support Schemes

New policies and support schemes (policy “think-tank” department)

• Advocating for Small Venues: we have shown that the loss of small venues in some areas were up to 60-70% in the world economic crisis. Our research shows what are the missing skills, regulatory problems, technology issues with re-opening these grassroots locations of the live music industry.

• Advocating for Better Taxation: our economic impact analysis software, which is partly open source, showed that in Slovakia, the music industry currently pays up to 100x more tax than the politically favored car manufacturing. Our model was also used to make calculations for the Hungarian film tax shelter.

• Advocating for Closing the Value Gap: we have calculated the euro value and the mechanism of value transfer from rightsholders to YouTube in two countries, and we are continuing with further countries.

Hungarian Music Industry Report — Slovak Music Industry Report — Made in Hungary

Copyright Credits

In this presentation Open data was used with open source software components. Detailed references in long-form publications for all software code used.

• The European statistical maps: © EuroGeographics for the administrative boundaries

• Visualization: © Daniel Antal, CEEMID, 2019.